EP0682414B1 - Dispositif d'estimation d'une valeur d'evaluation molle et dispositif d'estimation d'un systeme de vraisemblance maximale - Google Patents

Dispositif d'estimation d'une valeur d'evaluation molle et dispositif d'estimation d'un systeme de vraisemblance maximale Download PDF

Info

Publication number
EP0682414B1
EP0682414B1 EP94909296A EP94909296A EP0682414B1 EP 0682414 B1 EP0682414 B1 EP 0682414B1 EP 94909296 A EP94909296 A EP 94909296A EP 94909296 A EP94909296 A EP 94909296A EP 0682414 B1 EP0682414 B1 EP 0682414B1
Authority
EP
European Patent Office
Prior art keywords
metric
soft decision
output
symbol
estimation unit
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Expired - Lifetime
Application number
EP94909296A
Other languages
German (de)
English (en)
Other versions
EP0682414A4 (fr
EP0682414A1 (fr
Inventor
Shigeru Oki Electric Industry Co. Ltd. Ono
Hiroshi Oki Electric Industry Co. Ltd. Hayashi
Tomoko Oki Electric Industry Co. Ltd. Tanaka
Noriaki Oki Electric Industry Co. Ltd. Kondoh
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Oki Electric Industry Co Ltd
Original Assignee
Oki Electric Industry Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Oki Electric Industry Co Ltd filed Critical Oki Electric Industry Co Ltd
Publication of EP0682414A1 publication Critical patent/EP0682414A1/fr
Publication of EP0682414A4 publication Critical patent/EP0682414A4/fr
Application granted granted Critical
Publication of EP0682414B1 publication Critical patent/EP0682414B1/fr
Anticipated expiration legal-status Critical
Expired - Lifetime legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/03—Shaping networks in transmitter or receiver, e.g. adaptive shaping networks
    • H04L25/03006—Arrangements for removing intersymbol interference
    • H04L25/03178—Arrangements involving sequence estimation techniques
    • H04L25/03312—Arrangements specific to the provision of output signals
    • H04L25/03318—Provision of soft decisions
    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
    • H03M13/37—Decoding methods or techniques, not specific to the particular type of coding provided for in groups H03M13/03 - H03M13/35
    • H03M13/39—Sequence estimation, i.e. using statistical methods for the reconstruction of the original codes
    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
    • H03M13/37—Decoding methods or techniques, not specific to the particular type of coding provided for in groups H03M13/03 - H03M13/35
    • H03M13/39—Sequence estimation, i.e. using statistical methods for the reconstruction of the original codes
    • H03M13/41—Sequence estimation, i.e. using statistical methods for the reconstruction of the original codes using the Viterbi algorithm or Viterbi processors
    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
    • H03M13/63—Joint error correction and other techniques
    • H03M13/6331—Error control coding in combination with equalisation
    • H—ELECTRICITY
    • H03—ELECTRONIC CIRCUITRY
    • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
    • H03M13/00—Coding, decoding or code conversion, for error detection or error correction; Coding theory basic assumptions; Coding bounds; Error probability evaluation methods; Channel models; Simulation or testing of codes
    • H03M13/03—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words
    • H03M13/23—Error detection or forward error correction by redundancy in data representation, i.e. code words containing more digits than the source words using convolutional codes, e.g. unit memory codes

Definitions

  • the present invention relates to a receiver for digital communication. More in detail, it relates to a soft decision estimation unit for soft decision of the output of an adaptive equalizer to carry out soft decision decoding by a channel decoder in a high-speed digital communication receiver having the adaptive equalizer and the channel decoder. It also relates to a maximum-likelihood sequence estimation unit for estimating transmitted symbols.
  • an adaptive equalizer In order to correctly restore originally transmitted symbols from a received signal which has been affected by the fluctuation and noise, an adaptive equalizer is often used.
  • the adaptive equalizer of this kind the maximum-likelihood-sequence-estimation-type equalizer (referred to as an MLSE equalizer hereinafter) disclosed in the Citation 1 or a decision-feedback-type equalizer employed in the Citation 2 is often used.
  • a received and digitized signal is passed through a matched filter which minimizes the influence of noise by changing the characteristic thereof in accordance with the transmission path characteristic so that the most likely transmitted symbol sequence is estimated from the output of the matched filter.
  • Viterbi algorithm is often used as the maximum-likelihood estimation algorithm in this case as disclosed in the aforementioned Citation 3.
  • error-correction codes such as convolutional codes are used to reduce transmission data error. That is, in case the convolutional code is employed as the error-correction code, symbols formed by converting the transmission data into convolutional codes for transmission are modulated at the time of transmission and at the time of reception they are demodulated from the modulation frequency band to a base band and then are estimated by the adaptive equalizer. Thereafter the convolutional codes are decoded to restore the transmitted data.
  • Viterbi algorithm is usually used for decoding the convolutional codes.
  • Decoding the convolutional codes with the Viterbi algorithm is largely divided into two kinds, i.e., hard decision type and soft decision type, of which the soft decision type has better performance.
  • the output of the adaptive equalizer that is the input of the convolutional code decoding portion must be a soft decision.
  • the soft decision output type Viterbi algorithm As described in the Citation 1, in case the MLSE equalizer is used as the adaptive equalizer, it is necessary to employ the soft decision output type Viterbi algorithm as the maximum-likelihood estimation algorithm of the transmitted symbols in order to make the output of the adaptive equalizer a soft decision.
  • the soft decision output type Viterbi algorithm a quantity representing the certainty of its corresponding transmitted symbol (called as reliability) is updated every time a path representing the transmitted symbol is determined. Accordingly, it requires a reliability memory arranged in a matrix of number of states ⁇ number of transmitted symbols similarly to a path memory.
  • the decision feedback type equalizer employed in the Citation 2 is used as the adaptive equalizer receiver, it is possible to obtain a soft decision basically by outputting a signal before it is input to a decision unit in the adaptive equalizer.
  • the MLSE equalizer In the conventional soft decision estimation unit, however, in case the MLSE equalizer is used as the adaptive equalizer, it requires the reliability memory arranged in a matrix of number of states ⁇ number of transmitted symbols, and particularly when the maximum delay time of mutipath to be taken into consideration becomes long, the number of states which the soft decision output type Viterbi algorithm handles increases exponentially so that the capacity of the reliability memory becomes bulky. Moreover, there was also a problem that the amount of processes for calculating the reliability became bulky.
  • a soft decision corresponding to an estimated transmitted symbol does not simply correspond to each bit generally in a modulation system in which a symbol is represented by a plurality of bits such as QPSK and QAM which are used in high-speed digital mobile communication.
  • QPSK and QAM which are used in high-speed digital mobile communication.
  • deinterleaving must be achieved at the output of the adaptive equalizer, but there was a problem that is was difficult to do so with the soft decision as it is.
  • a noise-immune space diversity receiver having matched filters and a like number of channel estimators which are respectively coupled to diversity antennas to receive sequentially coded symbol sequences.
  • a branch metric calculator receives the outputs of the matched filters and the estimates from the channel estimators to calculate a branch metric of the received sequences for coupling to a maximum-likelihood estimator.
  • the branch metric is obtained by summing branch metric coefficients derived from channel estimates respectively with the output of the matched filters of by summing branch metric coefficients derived from a vector sum of channel estimates with the matched filter outputs.
  • adaptive channel estimators are provided for deriving channel estimates from received sequences and the output of an ML estimator.
  • First branch metrics are derived from the received sequences and supplied to a branch metric quality estimator in which quality estimates of the channels are derived from the first branch metrics.
  • An evaluation circuit evaluates the first branch metrics according to the quality estimates and produces a second branch metric for coupling to the ML estimator.
  • This known diversity receiver aims to communications systems in which the quality of the reception is significantly effected by channel noise and in the simple interference. It does not solve the problem of conventional solved decision estimation units as discussed with the citations 1-3 above.
  • a method of estimating a digitized data signal which signal is transmitted as a sequence of data symbols.
  • a near-instaneous soft decision is made for each estimated symbol.
  • Each state of a symbol at a given time increment the respective partial path metric is determined in respect of the state transition corresponding to the transmission of the first symbol type.
  • the respective partial path metric in respect to the state transition corresponding to the transmission of a second symbol type is determined.
  • the maximum value of the first and second symbol type partial path metric is selected, and the difference between the two maximum values at the preceding step is evaluated to yield an estimate of the transmitted data symbol in the form of a soft decision.
  • a soft decision estimation unit capable of estimating the soft decision with small capacity of memory and small amount of processing irrespective of the kind of an adaptive equalizer.
  • a soft decision estimation unit according to the invention is characterized in comprising the features of claim 1.
  • a maximum-likelihood sequence estimation unit comprises the features of claim 6.
  • Fig. 1 is a block diagram of a soft decision estimation unit according to a first embodiment of the invention
  • Fig. 2 is a block diagram showing an arrangement of a receiver employing the soft decision estimation unit
  • Fig. 3 is a schematic view showing an example of a burst signal employed by high-speed digital mobile communication
  • Fig. 4 is a block diagram of the soft decision estimation unit according to a second embodiment of the invention
  • Fig. 5 is a block diagram of the maximum-likelihood sequence estimation unit according to an embodiment of the invention
  • Fig. 6 is a block diagram of a soft decision estimation unit used in the maximum-likelihood sequence estimation unit according to the first embodiment of the invention
  • FIG. 7 is a block diagram showing an arrangement of a transmit/receive system of high-speed digital mobile communication employing the maximum-likelihood sequence estimation unit
  • Fig. 8 is a block diagram of a matched filter
  • Fig. 9 is a block diagram of the soft decision estimation unit used in the maximum-likelihood sequence estimation unit according to the second embodiment of the invention
  • Fig. 10 is a graph showing a computer simulation result of the maximum-likelihood sequence estimation unit according to the invention.
  • Fig. 1 is a block diagram of a soft decision estimation unit according to an embodiment of the invention.
  • an input terminal 1 is connected to the input end of a memory 3 and the output end of the memory 3 is connected to the input end of a transmission path estimating portion 6 and an input end of a metric calculating portion 8.
  • the output end of the transmission path estimating portion 6 is connected to another input end of the metric calculating portion 8.
  • an input terminal 2 is connected to the input end of a memory 4, and the output end of the memory 4 is connected to the input end of a symbol inverting portion 7.
  • the output end of the symbol inverting portion 7 is connected to still another input end of the metric calculating portion 8 and the output end of the metric calculating portion 8 is connected to the input end of a memory 5 and the input end of a subtracting portion 9.
  • the output end of the memory 5 is connected to the other input end of the subtracting portion 9.
  • the output end of the subtracting portion 9 is connected to the input end of a multiplying portion 10, and the output end of the multiplying portion 10 is connected to an output terminal 11.
  • the output end of the aforementioned memory 4 is further connected to the input end of the multiplying portion 10.
  • Fig. 2 is a block diagram showing an arrangement of a receiver employing the soft decision estimation unit according to the invention.
  • an input terminal 12 is connected to the input ends of an adaptive equalizer 13 and a soft decision estimation unit 14.
  • the output end of the adaptive equalizer 13 is also connected to another input end of the soft decision estimation unit 14.
  • the output end of the soft decision estimation unit 14 is connected to the input end of a channel decoding portion 15 and the output end of the channel decoding portion 15 is connected to an output terminal 16.
  • a signal received by the receiver is input to the adaptive equalizer 13 by way of the input terminal 12.
  • An antenna before the input terminal 12, a frequency converter, a decoder, various filters and the like after the input terminal 12 are omitted in Fig. 2 since they have no direct relationship with the present invention.
  • the adaptive equalizer 13 equalizes characteristics of the transmission path such as fading or multipath to estimate transmitted symbols.
  • various equalizers such as the MLSE equalizer or a decision feedback type equalizer can be used.
  • the output of the adaptive equalizer 13 is input to the soft decision estimation unit 14.
  • a signal output from the adaptive equalizer 13 is input to the input terminal 2 of the soft decision estimation unit 14 to be temporarily stored in the memory 4.
  • the signals stored in the memories 3 and 4 are usually digitized.
  • Fig. 3 shows a schematic arrangement of a burst signal employed by high-speed digital mobile communication, exemplifying a burst signal in GSM system that is a standard digital mobile communication system in Europe.
  • the soft decision estimation unit 14 performs processing by the burst, so that each of the memories 3 and 4 stores therein a burst of signal.
  • a signal stored in the memory 3 is supplied to the transmission path estimating portion 6 first to estimate the impulse response of the transmission path.
  • Many methods can be conceived for the estimation, for example in case of the GSM system employing the burst signal in Fig. 3, it is possible to comparatively easily perform the estimation using a training signal which is transmitted at about the central portion of the burst signal. That is, since the training signal is set to have an impulsive ideal auto-correlational characteristic, if it is generated in the transmission path estimating portion 6, a cross-correlation between the generated training signal and a signal corresponding to the training signal portion of the burst signal from the memory 3 represent the impulse response of the transmission path to be estimated.
  • the adaptive equalizer 13 is formed as the MLSE equalizer, since it has a transmission path estimating portion therein, the signal of the transmission path estimating portion in the adaptive equalizer 13 may be used without specially providing the transmission path estimating portion 6 in the soft decision estimation unit 14.
  • the signal stored in the memory 4 is supplied to the metric calculating portion 8 without being processed in the symbol inverting portion 7.
  • This signal is a sequence obtained by subjecting the transmitted signals to hard decision estimation, and the metric calculating portion 8 calculates a metric for the transmitted symbols, which were estimated at the adaptive equalizer 13 and then transmitted therefrom, by a method described later.
  • the metric is stored in a memory 5.
  • the adaptive equalizer 13 is formed as the MLSE equalizer, the metric is equal to the last metric in the adaptive equalizer 13, so that it can be directly supplied from the adaptive equalizer 13 too.
  • the symbol inverting portion 7 forms a symbol sequence by inverting the polarity of the first symbol of the transmitted symbol sequence stored in the memory 4 and supplies the same to the metric calculating portion 8.
  • the metric calculating portion 8 calculates a metric for the symbol sequence and supplies the result of calculation to the subtracting portion 9. Then, in the subtracting portion 9, the metric supplied from the metric calculating portion 8 is subtracted from the metric (i.e., a metric which is equal to the last metric in the adaptive equalizer 13, in case the adaptive equalizer 13 is formed as the MLSE equalizer) stored in the memory 5.
  • the output of the subtracting portion 9 is multiplied in the multiplying portion 10 by the first symbol of the transmitted symbol sequence which was estimated by the adaptive equalizer 13 and is stored in the memory 4 and the result of multiplication is output from the output terminal 11 as a first soft decision.
  • the symbol inverting portion 7 forms a symbol sequence by inverting the polarity of the second symbol of the transmitted symbol sequence stored in the memory 4 and supplies the same to the metric calculating portion 8.
  • the metric calculating portion 8 calculates a metric for the symbol sequence in which the second symbol has been inverted in polarity, and supplies the metric to the subtracting portion 9.
  • the metric supplied from the metric calculating portion 8 is subtracted from the metric stored in the memory 5.
  • the output of the subtracting portion 9 is multiplied in the multiplying portion 10 by the second symbol of the transmitted symbol sequence which is stored in the memory 4 and the result of multiplication is output from the output terminal 11 as a second soft decision. Thereafter the processing is repeated similarly until a soft decision for the last symbol is obtained.
  • the metric calculating portion 8 needs the content of the memory 3 that is a signal before it is input to the adaptive equalizer 13 and the impulse response of the transmission path that is the output of the transmission path estimating portion 6 for calculating a metric as described later, and they are supplied to the metric calculating portion 8 as illustrated in Fig. 1.
  • a sequence of soft decisions from the output terminal 11 of the soft decision estimation unit 14 is subjected to the soft decision decoding of the error correction codes in the channel decoding portion 15.
  • the soft decision type Viterbi algorithm can be used.
  • deinterleaving is also achieved in the channel decoding portion 15.
  • the output data of the channel decoding portion 15 is output from the output terminal 16.
  • the metric calculated in the metric calculating portion 8 will be described hereinafter in detail.
  • Various metrics can be conceived for use in the metric calculating portion 8, but the most basic one is a mean square error. It is a mean square error between the expected signal input to the adaptive equalizer 13 when a symbol sequence is transmitted (i.e., a signal obtained by calculating a convolution of the impulse response of the transmission path estimated by the transmission path estimating portion 6 and the symbol sequence which was used for calculating the metric) and a signal actually input to the adaptive equalizer 13. It is given by the following expression.
  • y(k) is the input signal to the adaptive equalizer 13 which is stored in the memory 3
  • h(m) is the impulse response of the transmission path estimated by the transmission path estimating portion 6
  • x(k) is a symbol sequence assumed for obtaining the metric
  • Es/N0 is an instant SNR.
  • the first ⁇ with respect to k means summing up as long as the length of the symbol sequence. This metric can be multiplied by a constant or a constant may be added thereto.
  • Fig. 4 is a block diagram of a soft decision estimation unit according to the present invention which is applied to such a case. The same elements as those in Fig. 1 are denoted at the same numerals.
  • the input terminal 1 is connected to the input end of the memory 3 while the output end of the memory 3 is connected to the input end of the transmission path estimating portion 6 and an input end of the metric calculating portion 8.
  • the output end of the transmission path estimating portion 6 is also connected to another input end of the metric calculating portion 8.
  • the input terminal 2 is connected to the input end of a binary-digitizing portion 17, and the output end of the binary-digitizing portion 17 is connected to the input end of the memory 4.
  • the output ends of the memory 4 are connected to the input ends of a bit inverting portion 18 and a unipolar/bipolar conversion portion 19 (referred to as a U/B conversion portion 19 hereinafter).
  • the output end of the bit inverting portion 18 is connected to the still another input end of the metric calculating portion 8, and the output ends of the metric calculating portion 8 is connected to the input end of the memory 5 and an input end of the subtracting portion 9.
  • the output end of the memory 5 is connected to the other input end of the subtracting portion 9.
  • the output end of the subtracting portion 9 is connected to an input end of the multiplying portion 10.
  • the other input end of the multiplying portion 10 is also connected to the output end of the aforementioned U/B conversion portion 19.
  • the output end of the multiplying portion 10 is connected to the output terminal 11.
  • a signal which is input to the adaptive equalizer 13, i.e., a signal input thereto from the input terminal 12 is supplied to the input terminal 1 and is temporarily stored in the memory 3.
  • the transmission path estimating portion 6 estimates the impulse response of the transmission path from the input signal of the adaptive equalizer 13 supplied thereto and supplies the result of estimation to the metric calculating portion 8 similarly to the case in Fig. 1.
  • a sequence which is obtained by subjecting the transmitted symbols output from the adaptive equalizer 13 to hard decision is supplied to the input terminal 2.
  • Each symbol of the transmitted symbol hard decision sequence is first converted into bits in the binary-digitizing portion 17.
  • the output bit sequence is stored in the memory 4.
  • the bit sequence stored in the memory 4 is supplied to the metric calculating portion 8 without being processed in the bit inverting portion 18, and a metric is calculated in the metric calculating portion 8 by a method, described later.
  • the metric is stored in the memory 5.
  • the bit inverting portion 18 forms a bit sequence from that stored in the memory 4 by inverting the first bit thereof and supplies the same to the metric calculating portion 8.
  • the metric calculating portion 8 calculates a metric corresponding to the bit sequence, and thereafter in the subtracting portion 9, the metric output from the metric calculating portion 8 is subtracted from the metric stored in the memory 5.
  • the output of the subtracting portion 9 is multiplied in the multiplying portion 10 by a value which is obtained by subjecting the first bit of the bit sequence stored in the memory 4 to (0,1) ⁇ (+1, -1) conversion in the U/B conversion portion.
  • the result of multiplication is output from the output terminal 11 as a first soft decision.
  • a bit sequence is formed by inverting the second bit of the bit sequence stored in the memory 4 and is supplied to the metric calculating portion 8.
  • the metric calculating portion 8 calculates a metric corresponding to the bit sequence in which the second bit is inverted.
  • the subtracting portion 9 the metric output from the metric calculating portion 8 is subtracted from the metric stored in the memory 5.
  • the output of the subtracting portion 9 is multiplied in the multiplying portion 10 by a value which is obtained by subjecting the second bit of the bit sequence stored in the memory 4 to (0,1) ⁇ (+1, -1) conversion in the U/B conversion portion.
  • the result of multiplication is output from the output terminal 11 as a second soft decision.
  • the processing is similarly performed until the last bit thereafter.
  • metric calculating portion 8 Various kinds of metrics to be used in the metric calculating portion 8 can be conceived similarly to the case of Fig. 1. That is, the metric used for Fig. 1 may be converted into a parameter which is related to the bit. In case a mean square error is used as the metric, it is possible to convert the bit sequence into a symbol sequence based on the modulation theory and calculate the metric using the expression (1)
  • FIGs. 1, 2 and 4 show functional blocks which constitutes the invention, each being realized by individual hardware, it is also possible to separately represent their functions by software using DSP etc.
  • Fig. 5 is a block diagram of the maximum-likelihood sequence estimation unit according to an embodiment of the invention.
  • an input terminal 21 is connected to the input end of a memory 22, and the output end of the memory 22 is connected to the input end of a transmission path estimating portion 25 and an input end of a matched filter 23.
  • the output end of the transmission path estimating portion 25 is connected to the input end of a coefficient setting portion 24, and the output end of the coefficient setting portion 24 is connected to the other input end of the matched filter 23.
  • the output end of the matched filter 23 is connected to the other input end of a memory 26, and the output end of the memory 26 is connected to an input end of a hard decision estimation unit 27.
  • the output end of the aforementioned transmission path estimating portion 25 is also connected to the other input end of the hard decision estimation unit 27.
  • the output end of the hard decision estimation unit 27 is connected to the input end of a soft decision estimation unit 28, and further the output ends of the transmission path estimating portion 25 and the memory 26 are connected to an input end of the soft decision estimation unit 28.
  • the output end of the soft decision estimation unit 28 is connected to an output terminal 29.
  • Fig. 6 is a block diagram of the soft decision estimation unit 28 which is used in the maximum-likelihood sequence estimation unit.
  • the same elements as those in Fig. 1 are denoted at the same numerals.
  • the input terminal 2 is connected to input ends of the memory 4.
  • the output end of the memory 4 is connected to the input end of the symbol inverting portion 7, and the output end of the symbol inverting portion 7 is connected to an input end of the metric calculating portion 8.
  • the output ends of the transmission path estimating portion 25 and the memory 26 illustrated in Fig. 5 are connected to the input ends of the metric calculating portion 8 by way of input terminals 14 and 15 respectively.
  • the output end of the metric calculating portion 8 is connected to the input ends of the memory 5 and the subtracting portion 9.
  • the output end of the memory 5 is connected to the other input end of the subtracting portion 9.
  • the output end of the subtracting portion 9 is connected to an input end of the multiplying portion 10, and the output end of the memory 4 is further connected to the other input end of the multiplying portion 10.
  • the output end of the multiplying portion 10 is connected to the output terminal 11.
  • Fig. 7 is a block diagram showing an arrangement of a transmit/receive system of high-speed digital mobile communication which employs the maximum-likelihood sequence estimation unit of the invention.
  • a transmission data input terminal 30 is connected to a coding portion 31.
  • the output end of the coding portion 31 is connected to the input end of a modulating portion 32, and the output end of the modulating portion 32 is connected to the input end of a frequency converter 34 by way of a transmission path 33.
  • the output end of the frequency converter 34 is connected to the input end of a low-pass filter (abbreviated as an LPF hereinafter) 35, and the output end of the LPF 35 is connected to the input end of an A/D converter 36.
  • LPF low-pass filter
  • the output end of the A/D converter 36 is connected to the input end of a maximum-likelihood sequence estimation unit 37, and the output end of the maximum-likelihood sequence estimation unit 37 is connected to the input end of a decoding portion 38.
  • the output end of the decoding portion 38 is connected to a data output terminal 39.
  • the maximum-likelihood sequence estimation unit 37 in Fig. 7 corresponds to the whole block diagram of Fig. 5.
  • the maximum-likelihood sequence estimation unit will be described in detail along the flow of signal with reference to Figs. 5 to 7.
  • a data is supplied from the transmission data input terminal 30 to the coding portion 31 in Fig. 7, an error correction coding is performed using, for example, convolutional codes.
  • the output of the coding portion 31 is modulated in the modulating portion 32 into a radio frequency band to be emitted as an electric wave.
  • the space is represented by the transmission path 33.
  • the signal received by the receiver is converted in frequency by the frequency converter 34 into a base-band signal by way of synchronous detection etc.
  • the modulating portion 32 employs a modulating system of orthogonal modulation type such as QPSK or MSK, it produces two outputs, i.e., the in-phase component and the quadrature phase component.
  • each signal processed after the frequency converter 34 has two components of the in-phase and the quadrature phase.
  • the signal output from the frequency converter 34 gets rid of noises which lie outside a desired frequency band in the LPF 35 and is converted into a digital signal in the A/D converter 36.
  • the digital signal output from the A/D converter 36 is supplied to the maximum-likelihood sequence estimation unit 37.
  • the maximum-likelihood sequence estimation unit 37 will be described further in detail hereinafter.
  • the digital signal supplied to the maximum-likelihood sequence estimation unit 37 is temporarily stored in the memory 22 by way of the input terminal 21 as shown in Fig. 5.
  • signals are generally transmitted forming bursts. Since the maximum-likelihood sequence estimation unit 37 processes the signal on a burst-by-burst basis, a burst of signal is stored in the memory 22.
  • the signal stored in the memory 22 is first to the transmission path estimating portion 25, where the impulse response of the transmission path is estimated.
  • Various methods of estimation can be conceived. For example, in case of a GSM system using a burst in Fig. 3, it is possible to estimate comparatively easily using a training signal which is transmitted around halfway through the burst. That is, since the training signal is designed to have an ideal impulsive autocorrelation characteristic, when the transmission path estimating portion 25 generates the training signal to get the correlation between the training signal and a signal corresponding thereto in the burst from the memory 22, it represents the impulse response of the transmission path to be estimated.
  • Fig. 8 is a block diagram showing an arrangement of the matched filter 23.
  • Fig. 8 shows a so-called transversal-type digital filter in which the order number N is 5.
  • a digital signal input through an input terminal 310 successively passes through delay portions 321, 322, ..., 32N each having a delay time which is equal to the sampling time T of the digital signal.
  • signals output from the input terminal 310 and the delay portions 321, 322, ..., 32N are multiplied by tap coefficients C0, C1, ..., CN in multiplying portions 330, 331, ..., 33N respectively.
  • Signals output from the multiplying portions 330, 331, ..., 33N are totaled in an adding portion 340 to be output from an output terminal 350.
  • a signal is usually represented by a complex number composed of a real part of in-phase component and an imaginary part of quadrature component for convenience. Therefore in this case, the tap coefficients C0, C1, ..., CN are complex numbers.
  • the tap coefficients C0, C1, ..., CN need to be set to the time-reversal complex conjugates of the impulse responses of the transmission path in order to function as matched filters to minimize the influence of noise.
  • the aforementioned setting is performed by the coefficient setting portion 24 which calculates the time-reversal complex conjugates of the impulse responses of the transmission path output from the transmission path estimating portion 25, and the time-reversal complex conjugates are supplied to the matched filter 23 as tap coefficients.
  • a signal output from the matched filter 23 is temporarily stored in the memory 26.
  • the signal stored in the memory 26 is supplied to the hard decision estimation unit 27 first, where it is subjected to the maximum-likelihood estimation of transmitted symbol sequence using a maximum-likelihood estimation algorithm.
  • the Viterbi algorithm disclosed in the Citation 1 can be employed as the maximum-likelihood estimation algorithm.
  • the Viterbi algorithm successively selects and determines most likely transmitted symbols using an estimated amount called metric, described later concretely. Calculating the metric requires the impulse response of the transmission path obtained by the transmission path estimating portion 25.
  • the signal output from the hard decision estimation unit 27 is supplied to the soft decision estimation unit 28, in which a soft decision is calculated.
  • the signal input to the soft decision estimation unit 28 is temporarily stored in the memory 4 by way of the input terminal 2 as shown in Fig. 6.
  • the signal stored in the memory 4, i.e., that input to the input terminal 2 is the hard decision of a transmitted symbol sequence estimated by the hard decision estimation unit 27 (estimated transmitted symbol sequence), and it is supplied first to the metric calculating portion 8 without being processed in the symbol inverting portion 7.
  • a metric corresponding to the input estimated transmitted symbol sequence is calculated.
  • the metric corresponding to the estimated transmitted symbol is, however, equal to the last metric in the hard decision estimation unit 27 so that it can be also directly supplied from the hard decision estimation unit 27.
  • the metric which is equal to the last metric calculated in the metric calculating portion 8 is stored in the memory 5.
  • the symbol inverting portion 7 forms a symbol sequence by inverting the polarity of the first symbol of the transmitted symbol sequence stored in the memory 4 and supplies the same to the metric calculating portion 8.
  • the metric calculating portion 8 calculates a metric corresponding to the symbol sequence in which the first symbol is inverted in polarity.
  • the metric stored in the memory 5 is supplied to the subtracting portion 9 and at the same time the metric calculated by the metric calculating portion 8 is also supplied to the subtracting portion 9.
  • the metric output from the metric calculating portion 8 is subtracted from the metric stored in the memory 5 (i.e., the metric which is equal to the last metric in the hard decision estimation unit 27).
  • the output of the subtracting portion 9 is multiplied in the multiplying portion 10 by the first symbol of the estimated transmitted symbol sequence stored in the memory 4, and the result of multiplication is output from the output terminal 11 as a first soft decision.
  • the symbol inverting portion 7 forms a symbol sequence by inverting the polarity of the second symbol of the estimated transmitted symbol sequence stored in the memory 4 and supplies the same to the metric calculating portion 8.
  • the metric calculating portion 8 calculates a metric corresponding to the symbol sequence in which the second symbol is inverted in polarity.
  • the subtracting portion 9 the metric output from the metric calculating portion 8 is subtracted from the metric stored in the memory 5.
  • the output of the subtracting portion 9 is multiplied in the multiplying portion 10 by the second symbol of the estimated transmitted symbol sequence stored in the memory 4 and the result of multiplication is output from the output terminal 11 as a second soft decision. Successively, similar processing is performed until a soft decision corresponding to the last symbol is obtained.
  • the metric calculating portion 8 requires the content of the memory 26 that is the output of the matched filter 23 and the impulse response of the transmission path that is the output of the transmission path estimating portion 25 in order to calculate the metrics described above.
  • the content of the memory 26 and the impulse response of the transmission path are supplied to the metric calculating portion 8 by way of the input terminals 14 and 15 respectively.
  • a soft decision sequence output from the output terminal 29 of the maximum-likelihood sequence estimation unit 37 is supplied to the decoding portion 38 to be subjected to the soft decision decoding of the error correcting codes. It can be performed by using a soft decision type Viterbi algorithm in case the convolutional codes are used as the error correcting codes.
  • Data output from the decoding portion 38 are output from the decoded data output terminal 39.
  • Various kinds of metrics can be conceived as those for use in the maximum-likelihood estimation algorithm of the hard decision estimation unit 27 or in the metric calculating portion 8 of the soft decision estimation unit 28, but most basic one is the mean square error. It is a mean square error between a signal actually input to the matched filter 23 and a signal obtained by the convolutional integration of the impulse response of the transmission path estimated by the transmission path estimating portion 25 and the symbol sequence which was used for calculating the metric. It can be expressed as follows.
  • y(k) is a signal input to the matched filter 23, i.e., the signal input to the maximum-likelihood sequence estimation unit 37 and stored in the memory 22
  • h(m) is the impulse response of the transmission path estimated by the transmission path estimating portion 25
  • x(k) is a symbol sequence used for calculating the metric
  • Es/NO is an instant SNR.
  • the first ⁇ with respect to k of the expression (1a) means summing up the following terms as long as the length of the symbol sequence while the latter ⁇ with respect to m means summing up so long as the impulse response of the transmission path h(m) has a value.
  • the expression (1a) represents a most basic metric
  • the matched filter 23 minimizes the influence of noise as described above, it is preferable to estimate the signal generally after the matched filter 23 rather than before the same. Accordingly, if a mean square error between the output signal of the matched filter 23 expected at the time of transmitting the symbol sequence and an actual output signal of the matched filter 23 is employed as a metric, it can be expressed as follows.
  • z(k) is a signal output from the matched filter 23.
  • s(m) which represents a response when the impulse response signal of the transmission path h(m) estimated by the transmission path estimating portion 25 passes through the matched filter 23 can be expressed as follows: wherein g(m) is the transfer function of the matched filter 23.
  • the transfer function g(m) of the matched filter 23 is the time-reversal complex conjugate of the impulse response h(m) of the transmission path
  • s(m) represents the autocorrelation of impulse response of the transmission path.
  • the first ⁇ with respect to k means summing up as long as the length of the symbol sequence
  • the latter ⁇ with respect to m means summing up the following terms so long as the autocorrelation s(m) of impulse response of the transmission path has a value
  • the ⁇ with respect to n means summing up as many as the number of taps of the matched filter 23. This metric may be multiplied by a constant or a constant may be added thereto.
  • Mean square errors represented by the above expressions (1a) and (2) are basic metrics, but they require complicated calculations, so that a quantity given by the following expression (4) which is a modification of the expressions (1a) and (2) is often used as the metric as described in the Citation 1.
  • Re [ ] means the real portion of the content inside [ ] and * means the complex conjugate.
  • ⁇ with respect to n represents summing up as long as the length of the assumed symbol sequence, while ⁇ s with respect to i and k represent summing up as long as the length of impulse response of the transmission path.
  • the expression (4) means subtracting the autocorrelation of cross-correlation between the symbol sequence and the impulse response of the transmission path from twice the real portion of the cross-correlation between the symbol sequence and the output of the matched filter 23.
  • the expression (4) can be also modified as follows:
  • the expression may be multiplied by a constant or a constant may be added to the expression.
  • Fig. 9 is a block diagram of the soft decision estimation unit 28 which is used in the maximum-likelihood sequence estimation unit according to another embodiment of the present invention and which uses the aforementioned expression (6) in particular.
  • the same elements as those in Fig. 6 are denoted at the same numerals.
  • a symbol sequence output from the hard decision estimation unit 27 is supplied to the input terminal 2.
  • the symbol sequence supplied to the input terminal 2 is temporarily stored in the memory 4 as it is.
  • the impulse response of the transmission path is input to an input terminal 15 from the transmission path estimating portion 25.
  • an autocorrelation calculating portion 41 calculates the autocorrelation (expression (3)) of impulse response of the transmission path input from the input terminal 15.
  • a summing portion 43 performs multiplying and summing operation aligning the first symbol of the symbol sequence stored in the memory 4 with the time origin of the autocorrelation function output from the autocorrelation calculating portion 41.
  • the output of the summing portion 43 is supplied to a subtracting portion 44.
  • the subtracting portion 44 is connected to an input terminal 14, through which the first output of the matched filter 23 is supplied to the subtracting portion 44. Then the subtracting portion 44 subtracts the output of the summing portion 43 from the first output of the matched filter 23.
  • the subtracting portion 44 supplies the result of subtraction to a multiplying portion 45.
  • a complex conjugate portion 42 supplies the complex conjugate of the first symbol read out from the aforementioned memory 4 to the multiplying portion 45, which multiplies the output of the complex conjugate portion 42 and that of the subtracting portion 44 together to supply the result of multiplication to a real portion selecting portion 46.
  • the real portion selecting portion 46 selectively outputs real numbers alone from the result of multiplication output from the multiplying portion 45.
  • the output of the real portion selecting portion 46 is output from the output terminal 11 as a first soft decision.
  • the second symbol of the symbol sequence stored in the memory 4 is subjected to similar processing to be output from the output terminal 11 as a second soft decision and similarly the following symbols are processed until the last symbol. Multiplying by a constant 4 as represented in the expression (6) may be performed by the multiplying portion 45 or may be omitted.
  • the impulse response of the transmission path estimated by the transmission path estimating portion 25 cannot be regarded to be constant.
  • the processing of the matched filter 23 may deteriorate performance instead of improving it.
  • the matched filter 23 had better do nothing or the output of the memory 22 in Fig. 5 had better supplied to the hard decision estimation unit 27 as it is.
  • calculating the soft decision which is the feature of the invention is effective. But in case the output of the memory 22 is input to the hard decision estimation unit 27 as it is, the output of the memory 22 is supplied to the soft decision estimation unit 28 too.
  • Fig. 10 is a graph exemplifying a computer simulation result of the maximum-likelihood sequence estimation unit according to the invention.
  • the axis of abscissa represents a ratio of signal power and noise power density per bit Eb/N0 and the axis of ordinate represents a bit error rate.
  • this case premises the condition of the GSM system.
  • Fig. 10 shows that the maximum-likelihood sequence estimation unit of the invention has approximately the same performance as a conventional one using a soft decision output type Viterbi algorithm.
  • the invention it is possible to estimate a soft decision from a transmitted symbol sequence or transmitted bit sequence which consists of hard decision outputs from an adaptive equalizer. As a result, it is possible to provide a soft decision estimation unit for estimating a soft decision with a small memory and a little amount of processing irrespective of the kind of the adaptive equalizer.
  • the maximum-likelihood sequence estimation unit of the invention can calculate a soft decision sequence after estimating a hard decision sequence without using the soft decision output type Viterbi algorithm. Furthermore, since this calculation is performed based on the hard decision sequence, the output of a matched filter and the impulse response of a transmission path, the maximum-likelihood sequence estimation unit does not need a reliability memory, so that it can output a soft decision with a small memory and a little amount of processing.

Landscapes

  • Physics & Mathematics (AREA)
  • Probability & Statistics with Applications (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Power Engineering (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Error Detection And Correction (AREA)
  • Cable Transmission Systems, Equalization Of Radio And Reduction Of Echo (AREA)

Claims (11)

  1. Unité d'estimation à décision douce, caractérisée en ce qu'elle comprend :
    un moyen de calcul de métrique (8) pour calculer successivement une première métrique correspondant à une séquence de symboles transmise estimée par un égaliseur adaptatif (13) et une seconde métrique correspondant à ladite séquence de symboles générée en inversant la polarité d'un symbole situé à une position à laquelle une décision douce est prise dans ladite séquence de symboles transmise ;
    un moyen de soustraction (9) pour soustraire ladite seconde métrique à ladite première métrique ; et
    un moyen de multiplication (10) pour multiplier ladite sortie dudit moyen de soustraction (9) par un symbole situé à une position à laquelle une décision douce est prise dans ladite séquence de symboles transmise.
  2. Unité d'estimation à décision douce selon la revendication 1, caractérisée en ce qu'elle comprend un moyen d'inversion de symbole (7) pour générer la séquence de symboles en inversant la polarité dudit symbole situé à une position à laquelle une décision douce est prise dans ladite séquence de symboles transmise.
  3. Unité d'estimation à décision douce selon la revendication 1 ou 2, caractérisée en ce que :
       ladite métrique est équivalente à une erreur quadratique moyenne entre un signal obtenu en calculant une convolution d'une réponse impulsionnelle d'un trajet de transmission et ladite séquence de symboles utilisée pour ledit calcul de métrique, ou est équivalente à ladite erreur quadratique moyenne multipliée par une constante ou à ladite erreur quadratique moyenne à laquelle une constante est ajoutée.
  4. Unité d'estimation à décision douce selon l'une quelconque des revendications 1 à 3, caractérisée en ce qu'elle comprend :
    un moyen de numérisation binaire (17) pour convertir un symbole transmis estimé par ledit égaliseur adaptatif (13) en une séquence de bits transmise en conformité avec un système de modulation ; et
    un moyen de conversion unipolaire/bipolaire (19) pour convertir un bit situé à une position à laquelle une décision douce est prise dans ladite séquence de bits, en une valeur bipolaire de ±1.
  5. Unité d'estimation à décision douce selon la revendication 4, caractérisée en ce qu'elle comprend :
       un moyen d'inversion de bit (18) pour générer une séquence de bits dans laquelle un bit situé à une position à laquelle une décision douce est prise dans ladite séquence de bits transmise est soumis à une inversion de polarité.
  6. Unité d'estimation de séquence à maximum de vraisemblance, comprenant :
    un moyen d'estimation de trajet de transmission (25) pour estimer une réponse impulsionnelle d'un trajet de transmission à partir d'un signal d'entrée ;
    un moyen d'estimation à décision dure (27) pour estimer une séquence de symboles transmise à maximum de vraisemblance sur la base dudit signal d'entrée en utilisant ladite réponse impulsionnelle dudit trajet de transmission ; et
    un moyen de calcul de décision douce (28) pour calculer une décision douce sur la base d'une sortie dudit moyen d'estimation à décision dure (27) en utilisant ledit signal d'entrée et ladite réponse impulsionnelle dudit trajet de transmission, caractérisé en ce que ledit moyen de calcul de décision douce (28) comprend :
    un moyen de calcul de métrique (8) pour calculer une première métrique correspondant à une séquence de symboles transmise estimée fournie en sorite par ledit moyen d'estimation à décision dure et une seconde métrique correspondant à ladite séquence de symboles générée en inversant la polarité d'un symbole situé à une position à laquelle une décision douce est prise dans ladite séquence de symboles transmise ;
    un moyen de soustraction (9) pour soustraire une métrique obtenue par ledit second moyen de calcul de métrique à ladite métrique obtenue par ledit premier moyen de calcul de métrique ; et
    un moyen de multiplication (10) pour multiplier ladite sortie dudit moyen de soustraction par un symbole situé à une position à laquelle une décision douce est prise dans ladite séquence de symboles transmise.
  7. Unité d'estimation de séquence à maximum de vraisemblance selon la revendication 6, caractérisée en ce qu'elle comprend :
       un moyen à filtre adapté (23) pour minimiser une influence du bruit dans ledit signal d'entrée au moyen de coefficients de prise ayant pour base ladite réponse impulsionnelle dudit trajet de transmission.
  8. Unité d'estimation de séquence à maximum de vraisemblance selon la revendication 6 ou 7, caractérisée en ce que ledit moyen de calcul de décision douce (28) comprend :
       un moyen d'inversion de symbole (7) pour générer la séquence de symboles dans laquelle un symbole se trouve à une position où une décision douce est prise dans ladite séquence de symboles transmise estimée fournie en sortie par ledit moyen d'estimation à décision dure (27) de polarité inversée.
  9. Unité d'estimation de séquence à maximum de vraisemblance selon la revendication 8, caractérisée en ce que ladite métrique est une erreur quadratique moyenne entre une intégrale de convolution de ladite réponse impulsionnelle d'un trajet de transmission et ladite séquence de symboles utilisée pour calculer une métrique, ladite intégrale de convolution ayant été amenée à passer à travers un filtre adapté et une sortie dudit moyen à filtre adapté (23), c'est-à-dire ladite métrique, est ladite erreur quadratique moyenne qui est multipliée par une constante ou à laquelle une constante est ajoutée.
  10. Unité d'estimation de séquence à maximum de vraisemblance selon l'une quelconque des revendications 6 à 9, caractérisée en ce que ladite métrique est une valeur obtenue en soustrayant une autocorrélation à une intercorrélation entre ladite séquence de symboles utilisée pour calculer ladite métrique et ladite réponse impulsionnelle d'un trajet de transmission, à une valeur obtenue en doublant la partie réelle de l'intercorrélation entre une séquence de symboles estimée par ledit moyen d'estimation à décision dure (27) et la sortie dudit moyen à filtre adapté (23), ou bien ladite métrique est ladite valeur calculée qui est multipliée par une constante ou à laquelle est ajoutée une constante.
  11. Unité d'estimation de séquence à maximum de vraisemblance selon l'une quelconque des revendications 6 à 10, caractérisée en ce qu'elle comprend :
    un moyen de calcul d'autocorrélation (41) pour calculer une autocorrélation de ladite réponse impulsionnelle dudit trajet de transmission ;
    un moyen de sommation (43) pour effectuer l'opération de multiplication et de sommation en alignant un symbole situé à une position à laquelle une décision douce est prise dans une séquence de symboles transmise estimée par ledit moyen d'estimation à décision dure (27) avec une origine des temps de ladite autocorrélation qui est calculée par ledit moyen de calcul d'autocorrélation (41) ;
    un moyen de soustraction (44) pour soustraire une sortie dudit moyen de sommation (43) à une sortie dudit moyen à filtre adapté (23) à une position correspondant à un symbole pour lequel une décision douce est prise ;
    un moyen de conjugaison complexe (42) pour obtenir un conjugué complexe dudit symbole pour lequel une décision douce est prise ;
    un moyen de multiplication (45) pour multiplier une sortie dudit moyen de soustraction (44) par une sortie dudit moyen de conjugaison complexe (42) ; et
    un moyen de sélection de partie réelle (46) pour obtenir la partie réelle d'une sortie dudit moyen de multiplication (45).
EP94909296A 1993-11-29 1994-03-10 Dispositif d'estimation d'une valeur d'evaluation molle et dispositif d'estimation d'un systeme de vraisemblance maximale Expired - Lifetime EP0682414B1 (fr)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
JP29784793 1993-11-29
JP297847/93 1993-11-29
JP29784793 1993-11-29
PCT/JP1994/000386 WO1995015034A1 (fr) 1993-11-29 1994-03-10 Dispositif d'estimation d'une valeur d'evaluation molle et dispositif d'estimation d'un systeme de vraisemblance maximale

Publications (3)

Publication Number Publication Date
EP0682414A1 EP0682414A1 (fr) 1995-11-15
EP0682414A4 EP0682414A4 (fr) 1997-02-26
EP0682414B1 true EP0682414B1 (fr) 2001-11-21

Family

ID=17851925

Family Applications (1)

Application Number Title Priority Date Filing Date
EP94909296A Expired - Lifetime EP0682414B1 (fr) 1993-11-29 1994-03-10 Dispositif d'estimation d'une valeur d'evaluation molle et dispositif d'estimation d'un systeme de vraisemblance maximale

Country Status (5)

Country Link
US (1) US6302576B1 (fr)
EP (1) EP0682414B1 (fr)
DE (1) DE69429161T2 (fr)
FI (1) FI953461A (fr)
WO (1) WO1995015034A1 (fr)

Families Citing this family (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5805637A (en) * 1995-09-29 1998-09-08 Sony Corporation Automatic equalizer and digital signal reproducing apparatus carrying the same
US5905757A (en) * 1996-10-04 1999-05-18 Motorola, Inc. Filter co-processor
US6614860B1 (en) * 1997-02-04 2003-09-02 Nokia Telecommunications Oy Compensation of doppler shift in a mobile communication system
FI102335B (fi) * 1997-02-28 1998-11-13 Nokia Telecommunications Oy Vastaanottomenetelmä ja vastaanotin
EP0945989A1 (fr) * 1998-03-12 1999-09-29 Hitachi Micro Systems Europe Limited Décodage Viterbi
JP3230482B2 (ja) * 1998-03-13 2001-11-19 日本電気株式会社 適応等化器
US6185251B1 (en) * 1998-03-27 2001-02-06 Telefonaktiebolaget Lm Ericsson Equalizer for use in multi-carrier modulation systems
WO2000001124A1 (fr) * 1998-06-29 2000-01-06 Nokia Networks Oy Estimation de symboles a l'aide d'un algorithme et d'une retropropagation de sortie ponderee
US6529567B1 (en) * 1999-03-10 2003-03-04 Motorola, Inc. Electronic apparatus and method for receiving noisy signals
US6466616B1 (en) * 1999-07-02 2002-10-15 Telefonaktiebolaget Lm Ericsson (Publ) Power efficient equalization
DE19935824A1 (de) * 1999-07-29 2001-02-15 Siemens Ag Verfahren zum Erzeugen von Zuverlässigkeitsinformationen für die Kanaldecodierung in einem Funkempfänger sowie entsprechender Funkempfänger
US6625236B1 (en) 2000-02-08 2003-09-23 Ericsson Inc. Methods and systems for decoding symbols by combining matched-filtered samples with hard symbol decisions
US7058141B1 (en) * 2000-06-02 2006-06-06 Nec Usa, Inc. MLSE decoding of PRS type inter-bin interference in receiver-end windowed DMT system
US6708308B2 (en) * 2001-01-10 2004-03-16 International Business Machines Corporation Soft output viterbi algorithm (SOVA) with error filters
US6823027B2 (en) * 2001-03-05 2004-11-23 Telefonaktiebolaget Lm Ericsson (Publ) Method for enhancing soft-value information
US7453933B2 (en) * 2001-05-04 2008-11-18 Lucent Technologies Inc. Method of estimating a signal-to-interference+noise ratio (SINR) using data samples
KR100450949B1 (ko) * 2001-09-18 2004-10-02 삼성전자주식회사 데이터 통신 시스템에서 채널 복호기 입력 연성 결정 값계산 장치 및 방법
US7107514B1 (en) 2002-08-27 2006-09-12 Marvell International Ltd. Method and apparatus for detecting Viterbi decoder errors due to quasi-catastophic sequences
US8144572B2 (en) * 2004-09-14 2012-03-27 Qualcomm Incorporated Detection and mitigation of interference and jammers in an OFDM system
US7965782B1 (en) 2007-05-11 2011-06-21 Redpine Signals, Inc. Reduced complexity maximum likelihood decoder for MIMO communications
JP5299130B2 (ja) * 2009-07-03 2013-09-25 富士通セミコンダクター株式会社 受信データ処理回路及び受信データ処理切り替え方法
CN111566953B (zh) * 2018-01-19 2022-11-01 日本电信电话株式会社 符号判定装置和符号判定方法

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US4571734A (en) * 1983-08-05 1986-02-18 International Business Machines Corporation Method and apparatus for decoding the output signal of a partial-response class-IV communication or recording-device channel
JPS61161027A (ja) * 1985-01-08 1986-07-21 Mitsubishi Electric Corp ヴイタビ復号器
JPS61230430A (ja) * 1985-04-04 1986-10-14 Mitsubishi Electric Corp 加算比較選択回路
US5144644A (en) * 1989-10-13 1992-09-01 Motorola, Inc. Soft trellis decoding
US5271042A (en) * 1989-10-13 1993-12-14 Motorola, Inc. Soft decision decoding with channel equalization
GB2238692B (en) 1989-11-29 1993-12-01 Technophone Ltd Data symbol estimation
JP2614524B2 (ja) * 1990-03-02 1997-05-28 国際電信電話株式会社 誤り訂正符号の復号方法
EP0449327B1 (fr) * 1990-03-30 1998-07-15 Nec Corporation Récepteur à diversité d'espace et à immunité contre le bruit
GB2246272B (en) * 1990-07-19 1994-09-14 Technophone Ltd Maximum likelihood sequence detector
US5134635A (en) 1990-07-30 1992-07-28 Motorola, Inc. Convolutional decoder using soft-decision decoding with channel state information
US5142551A (en) * 1991-02-28 1992-08-25 Motorola, Inc. Signal weighting system for digital receiver
US5214675A (en) 1991-07-02 1993-05-25 Motorola, Inc. System and method for calculating channel gain and noise variance of a communication channel
US5414738A (en) * 1993-11-09 1995-05-09 Motorola, Inc. Maximum likelihood paths comparison decoder
US5471500A (en) * 1994-03-08 1995-11-28 At&T Ipm Corp. Soft symbol decoding

Also Published As

Publication number Publication date
WO1995015034A1 (fr) 1995-06-01
FI953461A (fi) 1995-09-25
DE69429161D1 (de) 2002-01-03
DE69429161T2 (de) 2002-06-20
FI953461A0 (fi) 1995-07-18
EP0682414A4 (fr) 1997-02-26
US6302576B1 (en) 2001-10-16
EP0682414A1 (fr) 1995-11-15

Similar Documents

Publication Publication Date Title
EP0682414B1 (fr) Dispositif d'estimation d'une valeur d'evaluation molle et dispositif d'estimation d'un systeme de vraisemblance maximale
US6690739B1 (en) Method for intersymbol interference compensation
US6625236B1 (en) Methods and systems for decoding symbols by combining matched-filtered samples with hard symbol decisions
EP1264456B1 (fr) Procede et appareil assurant a la fois l'elimination de l'interference fondee sur la decision ponderee et le decodage
US5621764A (en) Soft decision signal outputting receiver
AU728540B2 (en) Method and apparatus for digital symbol detection using transmission medium response estimates
US6937648B2 (en) Equalizer for communication over noisy channels
US7248647B2 (en) Radio telecommunications system operative by interactive determination of soft estimates, and a corresponding method
US6269124B1 (en) Data transmission system, receiver, and recording medium
US7298778B2 (en) Sub-optimal iterative receiver method and system for a high-bit-rate CDMA transmission system
US6990142B2 (en) Process and device for estimating the impulse response of an information transmission channel, in particular for a cellular mobile telephone
EP0731567B1 (fr) Estimateur pour évaluer la séquence de la probabilité maximale et procédé pour la mise en oeuvre
EP1217798B1 (fr) Procédé d'égalisation adaptatif et égaliseur adaptatif
US20030115061A1 (en) MPSK equalizer
US6724841B2 (en) Equalizer with a cost function taking into account noise energy
Hart et al. Maximum-likelihood synchronization, equalization, and sequence estimation for unknown time-varying frequency-selective Rician channels
JP3424723B2 (ja) 適応等化器
JP4401048B2 (ja) 情報送信チャネルのインパルス応答を推定する方法
US7035357B2 (en) Process and device for extracting digital data contained in a signal conveyed by an information transmission channel, in particular for a cellular mobile telephone
JP2001525139A (ja) 周波数選択デジタル伝送チャネルの等化及び復号装置
JP2786335B2 (ja) 軟判定値推定器および最尤系列推定器
Wu et al. A two-stage MSK-type detector for low-BT GMSK signals
Hart et al. Extended MLSE receiver for the frequency-flat, fast-fading channel
Femenias et al. Dual MRC diversity reception of TCM-MPSK signals over Nakagami fading channels
Dong Turbo equalization with channel prediction and iterative channel estimation

Legal Events

Date Code Title Description
PUAI Public reference made under article 153(3) epc to a published international application that has entered the european phase

Free format text: ORIGINAL CODE: 0009012

17P Request for examination filed

Effective date: 19950726

AK Designated contracting states

Kind code of ref document: A1

Designated state(s): DE FR GB IT SE

A4 Supplementary search report drawn up and despatched
AK Designated contracting states

Kind code of ref document: A4

Designated state(s): DE FR GB IT SE

17Q First examination report despatched

Effective date: 20000225

GRAG Despatch of communication of intention to grant

Free format text: ORIGINAL CODE: EPIDOS AGRA

GRAG Despatch of communication of intention to grant

Free format text: ORIGINAL CODE: EPIDOS AGRA

GRAH Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOS IGRA

GRAH Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOS IGRA

GRAH Despatch of communication of intention to grant a patent

Free format text: ORIGINAL CODE: EPIDOS IGRA

GRAA (expected) grant

Free format text: ORIGINAL CODE: 0009210

RIC1 Information provided on ipc code assigned before grant

Free format text: 7H 03M 13/23 A, 7H 04L 25/30 B, 7H 04L 27/00 B

AK Designated contracting states

Kind code of ref document: B1

Designated state(s): DE FR GB IT SE

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: IT

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT;WARNING: LAPSES OF ITALIAN PATENTS WITH EFFECTIVE DATE BEFORE 2007 MAY HAVE OCCURRED AT ANY TIME BEFORE 2007. THE CORRECT EFFECTIVE DATE MAY BE DIFFERENT FROM THE ONE RECORDED.

Effective date: 20011121

REG Reference to a national code

Ref country code: GB

Ref legal event code: IF02

REF Corresponds to:

Ref document number: 69429161

Country of ref document: DE

Date of ref document: 20020103

ET Fr: translation filed
PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: SE

Free format text: LAPSE BECAUSE OF FAILURE TO SUBMIT A TRANSLATION OF THE DESCRIPTION OR TO PAY THE FEE WITHIN THE PRESCRIBED TIME-LIMIT

Effective date: 20020221

PLBE No opposition filed within time limit

Free format text: ORIGINAL CODE: 0009261

STAA Information on the status of an ep patent application or granted ep patent

Free format text: STATUS: NO OPPOSITION FILED WITHIN TIME LIMIT

26N No opposition filed
REG Reference to a national code

Ref country code: GB

Ref legal event code: 732E

REG Reference to a national code

Ref country code: FR

Ref legal event code: TP

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: DE

Payment date: 20130331

Year of fee payment: 20

Ref country code: GB

Payment date: 20130320

Year of fee payment: 20

PGFP Annual fee paid to national office [announced via postgrant information from national office to epo]

Ref country code: FR

Payment date: 20130417

Year of fee payment: 20

REG Reference to a national code

Ref country code: DE

Ref legal event code: R071

Ref document number: 69429161

Country of ref document: DE

REG Reference to a national code

Ref country code: GB

Ref legal event code: PE20

Expiry date: 20140309

PG25 Lapsed in a contracting state [announced via postgrant information from national office to epo]

Ref country code: DE

Free format text: LAPSE BECAUSE OF EXPIRATION OF PROTECTION

Effective date: 20140311

Ref country code: GB

Free format text: LAPSE BECAUSE OF EXPIRATION OF PROTECTION

Effective date: 20140309